Accel 2025 Globalscape: Race for Compute

2025
|
Globalscape
Race for
Compute
NOVEMBER 2025

Legals and disclosures
Accel Confidential
GLOBALSCAPE REPORT | NOVEMBER 2025
GLOBALSCAPE
2
This document has been prepared by Accel (including its UK-based investment advisor Accel Partners Management LLP (“Accel London”)). Accel London is a
limited liability partnership incorporated in England and Wales with registered number OC303424, having its registered office at 1 New Burlington Place,
London, W1S 2HR. Accel London is authorised and regulated by the Financial Conduct Authority with registration number 222503.
This document is not directed at and should not be distributed, published, or used in any jurisdiction where such distribution, publication, or use would
be unlawful.
This document is provided for information purposes only and does not constitute an offer, or a solicitation or recommendation, to purchase or sell any
security, investment product or service and should not be treated as such. This document forms part of a presentation to persons who have been invited to
attend such presentation and must be read and considered in the context of that presentation only.
This document is the property of Accel and is addressed to its recipients exclusively and confidentially for their personal use. It may not be reproduced
(in whole or in part), transmitted, modified, or used for any other purpose without Accel’s prior written permission.
This document does not contain personalised recommendations or advice, is not intended to endorse or recommend a particular course of action or to be
construed or relied on as investment, legal, business, accounting, tax or other professional advice. It does not substitute professional advice on
investment in financial products. As such, before entering into any transaction, recipients should consider carefully the suitability of a transaction to
their particular circumstances and, where necessary, obtain independent professional advice in respect of risks, as well as any legal, regulatory,
credit, tax, and accounting consequences.
All information used in this document has been obtained from sources believed to be reliable as at their date of publication, but accuracy cannot be
guaranteed. No representation or warranties are given or implied with respect to information contained in this document and no reliance should be placed
on the contents of this document.

01 — INTRODUCTION 02 — NEW WORLD 03 — 21ST CENTURY UTILITY 04 — 2025 GLOBALSCAPE 05 — WHAT’S NEXT 06 — APPENDIX Accel Confidential 3
NOVEMBER 2025
GLOBALSCAPE
01
Introduction
02
The new world of software
03
AI as the 21st century utility
04
Accel 2025 Globalscape
05
What’s next
06
Appendix
2025
|
Globalscape

We are at the beginning of a new
industrial revolution … over the
course of the next four or five years
we’ll have $2T worth of data centers
that will be powering software
around the world.
INTRODUCTION
4
— JENSEN HUANG | CEO OF NVIDIA

Source: CapIQ
Notes: Comparison between 29 October as of every calendar year. Data as of Oct-29, 2025
0
7500
15000
22500
30000
Oct-2010 Oct-2011 Oct-2012 Oct-2013 Oct-2014 Oct-2015 Oct-2016 Oct-2017 Oct-2018 Oct-2019 Oct-2020 Oct-2021 Oct-2022 Oct-2023 Oct-2024 Oct-2025
AI is pushing NASDAQ to an all-time high, at a similar
pace to earlier platform shifts…
INTRODUCTION
5
2.5K
5.1K
24.0K
MOBILE
+102% 2010-2015
NASDAQ Composite Index (Oct-2010 / Oct-2025)
CLOUD
+120% 2015-2020
AI
+114% 2020-2025
11.2K

…not impacted by geopolitical
and macro uncertainty
INTRODUCTION
6

AI is concentrating tech market leadership
INTRODUCTION
7
Source: CapIQ, Accel analysis
Notes: Data as of Oct-29 2025
“Super Six” companies accounting for ~50% of the NASDAQ market cap
NASDAQ Composite Market Cap Super Six Financial Overview
$20.7T
Combined Market
Cap
$1.9T
Combined
CY24 Rev
30%
Avg. CY24 Rev
Growth (y/y)
$0.6T
Combined CY24
Op. Cash Flow
Other, $21.0T
Alphabet, $3.3T Amazon, $2.5T
Meta, $1.9T
Apple, $4.0T
50%
Nvidia, $5.0T
12%
Microsoft, $4.0T
10%
10%
8% 6%
4%

“Super Six” added $4.9T of market cap in the past year
INTRODUCTION
8
Market Capitalisation (Oct 24 vs. Oct 25)
$1.9T
$2.5T
$3.3T
$4.0T
$4.0T
$5.0T
$1.5T
$2T
$2.1T
$3.6T
$3.2T
$3.5T
Oct-2024 Oct-2025
Notes: Comparison between Oct-29 2024 and Oct-29 2025. Data as of Oct-29 2025
+$1.6T +$0.8T +$0.5T +$1.2T +$0.5T +$0.4T

Note: “Lifted by AI” includes >$100B enterprise software companies with market capitalization growth >10% y/y. “Still waiting for AI adoption in enterprise” includes
>$100B enterprise software companies with negative market capitalization growth y/y. Comparison between Oct-29 2024 and Oct-29 2025. Data as of Oct-29 2025
Mixed story for enterprise cloud giants as agentic
adoption is still in its early days
Accel Confidential
INTRODUCTION
9
CLOUD COMPANIES WITH MARKET CAP >$100B
Oracle Palantir IBM Shopify Palo
Alto Networks
CrowdStrike
$137B
$147B
$233B
$288B
$472B
$785B
$76B
$120B
$104B
$194B
$101B
$481B
Still waiting for AI
adoption in enterprise
Oct-2024 Oct-2025
ServiceNow Salesforce Adobe
$141B
$239B
$189B
$214B
$286B
$197B
Lifted by AI
AI Infra
Bridging raw
models to
business use-
cases
AI-powered
commerce
AI distribution
to enterprise
AI cyber threats

Note: Selection of US, European and Israeli cloud companies (including previously public companies). Companies that comprise the Accel Globalscape Public Cloud Index are
listed in the appendix. Data as of Oct-29 2025
Source: Accel analysis, CapIQ
Software is not going away: Globalscape Public Cloud
Index recovering (+25% y/y)
Accel Confidential
INTRODUCTION
10
Globalscape Public Cloud Index
0
100
200
300
400
500
600
700
800
900
1000
2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025
+25% y/y

Note: Selection of US, European and Israeli cloud companies (including previously public companies). Companies that comprise the Accel Globalscape Public Cloud Index are
listed in the appendix. Data as of Oct-29 2025
Source: Accel analysis, CapIQ
0.0x
6.0x
12.0x
18.0x
24.0x
D
e
c
-
0
9
M
a
y
-
1
0
O
c
t
-
1
0
M
a
r
-
1
1
A
u
g
-
1
1
J
a
n
-
1
2
J
u
n
-
1
2
N
o
v
-
1
2
A
p
r
-
1
3
S
e
p
-
1
3
F
e
b
-
1
4
J
u
l
-
1
4
D
e
c
-
1
4
M
a
y
-
1
5
O
c
t
-
1
5
M
a
r
-
1
6
A
u
g
-
1
6
J
a
n
-
1
7
J
u
n
-
1
7
N
o
v
-
1
7
A
p
r
-
1
8
S
e
p
-
1
8
F
e
b
-
1
9
J
u
l
-
1
9
D
e
c
-
1
9
M
a
y
-
2
0
O
c
t
-
2
0
M
a
r
-
2
1
A
u
g
-
2
1
J
a
n
-
2
2
J
u
n
-
2
2
N
o
v
-
2
2
A
p
r
-
2
3
S
e
p
-
2
3
F
e
b
-
2
4
J
u
l
-
2
4
D
e
c
-
2
4
M
a
y
-
2
5
O
c
t
-
2
5
Multiples are now back above historical pre-COVID levels
Accel Confidential
INTRODUCTION
11
Globalscape Public Cloud Index - EV / NTM Rev
7.8x
Oct-25
7.1x
Pre-COVID Avg

Tech IPO markets starting to reopen
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025
8
4
1
0
46
19
13
12
10
7
12
15
11
9
4
5
Source: Accel Analysis, Qatalyst
INTRODUCTION
12
Software & AI IPOs by year

Source: Accel data refers to SaaS, Cloud and AI investments. Data as of Oct-22 2025
Accel is investing in software & AI globally
INTRODUCTION
13
23+
IPOs
80+
UNICORNS
450+
COMPANIES
$13B+
INVESTED
Selected Accel investments in the last 24 months

2025
|
Globalscape
NOVEMBER 2025
GLOBALSCAPE
01
Introduction
02
03
04
05
What’s next
06
Appendix

Q1-21 Q2-21 Q3-21 Q4-21 Q1-22 Q2-22 Q3-22 Q4-22 Q1-23 Q2-23 Q3-23 Q4-23 Q1-24 Q2-24 Q3-24 Q4-24 Q1-25 Q2-25 Q3-25
15%
17%
16%
16%
16%
17%
18%
19%
20%
21%
27%
32%
35%
39%
45%
45%
45%
47%
42%
Public cloud companies’ growth rates continue
to be under pressure
Accel Confidential
THE NEW WORLD OF SOFTWARE
15
Average annual growth rates by quarter - Globalscape Cloud Index
Note: Growth rate is most recent quarter revenue growth (YoY). Companies that comprise the Accel Globalscape Public Cloud Index are listed in the appendix. Data as of
Oct-22 2025

Notes: 2025 Annualisation Excludes deals >$20B EV; Chart only includes deals >$500M EV
Source: Accel Analysis, Qatalyst
Sponsors activity continues to be
strong for legacy software companies
Accel Confidential
16
AGGREGATE SPONSOR-LED M&A FOR DEALS >$500M
2018 2019 2020 2021 2022 2023 2024 2025E
$19B
$35B
$48B
$45B
$61B
$31B
$10B
$21B $22B
$13B
$4B
$1B
$23B
$7B
$1B
$6B
$75B
$61B
$49B
$61B
$54B
$17B
$22B
$6B
Private Buyouts
Take-Privates
Annualization
LARGEST SPONSOR-LED M&A BY ENTERPRISE VALUE
Target EV Acquirer
$12.3B
(Digital Aviation) $10.6B
$6.3B
$5.4B
$5.3B
$4.5B
$3.0B
$2.6B

M&A is accelerating driven by consolidation
and big strategic initiatives
17
AGGREGATE STRATEGIC M&A FOR DEALS >$500M
2018 2019 2020 2021 2022 2023 2024 2025E
$23B
$150B
$80B
$57B
$67B
$51B
$12B
$19B
$41B
LARGEST STRATEGIC M&A BY EV LTM
Target EV Acquirer
$32.0B
$26.2B
$17.0B
$8.2B
$5.1B
$4.2B
$3.2B
$3.0B
Notes: 2025 Annualisation Excludes deals >$20B EV; Chart only includes deals >$500M EV
Source: Accel analysis, Qatalyst

Source: JP Morgan Research, Salesforce, Microsoft Q3-25 Earnings Call, Atlassian Q4 FY25 & Q1 FY26 Earnings
Public software companies are investing
heavily in agentic capabilities
18
Salesforce Agentforce Microsoft Copilot Studio Atlassian
Released November 2025
~$440M ‘Agentic’ ARR
~13K Customers
230K B2B Users
>1M Agents Created (Q3-25)
1.5K+ out-of-the-box connectors
$610M acquisition of The Browser Company
3.5M MAUs of AI Tools (Q1 FY26)
5x QoQ Token Usage (Q4 FY25)

Source: UiPath, Decagon, n8n
Early enterprise agentic use cases are promising
19
Decagon x Duolingo n8n x Vodafone
By streamlining DET’s customer
support operations, Decagon
enables Duolingo to focus on what
matters most — providing countless
test-takers each year with a
seamless, accessible, and stress-
free experience.
— Ian Riggins, Senior Operations
Manager for Duolingo English Test
[n8n] provides SOAR capability and
a workflow capability in a low-
code model … it did everything that
we wanted, all in one tool … to
manually carry out this process …
would have been impossible
without a significant increase in
operational resources.
— Claire Van Hinsbergh,
Engineering Manager at Vodafone
SELECTED CASE STUDIES
UiPath x Fiserv
• 98% end-to-end automation
• 12K hrs saved
• 7+ processes automated
By partnering with UiPath, we’ve
been able to operationalise AI
safely and effectively, while
showcasing its enterprise-wide
potential through key use cases.
— Sharbs Shaaya, Senior Director -
AI Operations at Fiserv
UiPath robots extract merchant details from
the database, cross-check them via a Bing API,
and select the correct MCC codes using
generative AI prompts from UiPath GenAI
Activities.
Implementing Celonis AI Assistant
for Credit Block Management has
been a game changer for our
order management operations,
streamlining our processes and
resulting in faster, more reliable
outcomes.
— Rafael Domene, Global CIO at
Cosentino
Celonis x Cosentino
• 80% ticket deflection rate
• Automatic syncing with FAQs
• 500M users
• £3.6M annualized cost savings
• 33 security-focused workflows
• 5K person-days saved
• >1.8K daily blocked orders
reviewed by AI
• 5 min saved per credit
block review
• Up to 28 day expected order
cycle time reduction

Source: Public information and Accel analysis
A new breed of private AI native applications
shows unprecedented growth…
20
$ARR
$0M
$25M
$50M
$75M
$100M
Years
1 2 3 4 5 6
TIME FROM $1→$100M ARR
Years

2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025E
$15B
$19
$47B
$29B
$39B
$16B
$2B
$2B
$0B
$1B
$0B
$0B
$1B
$74B
$64B
$47B
$127B
$181B
$70B
$57B
$46B
$30B
$25B
$26B
$184B
$103B
$63B
$129B
$183B
$70B
$58B
$46B
$30B
$27B $25B
Cloud & AI Application Funding (ex. Models)
AI Model Funding
OpenAI
Anthropic
xAI
Source: Dealroom, Accel analysis
Note: Data as of Oct-22 2025. 2025 data is annualized.
…driving venture funding to an all-time high
Accel Confidential
21
TOTAL EU/IL/US VENTURE CAPITAL IN CLOUD & AI, $B
3.7% 0.0% 0.1% 0.0% 1.8% 0.1% 1.3% 1.3% 24.9% 37.8% 59.6%
Model funding as % of total
$15B
$19B

AI Model funding is heavily concentrated
in the US vs EU/IL…
22
2021 2022 2023 2024 2025E
$4B
$2B
$1B
$0B
$0B
$106B
$36B
$15B
$1B
$2B
US Model Funding
EU/IL Model Funding
US VS. EU/IL VENTURE CAPITAL AI MODEL INVESTMENTS, $B
Source: Dealroom, Pitchbook, Accel analysis

Company Founded Last Round
Total
Funding
Valuation
2023 $1.5B $2.7B $13.7B
2021 $579M $579M $1.8B
2017 $300M $637M $1.5B
2022 $250M $350M $6.6B
2019 $213M $253M $1.2B
…driven by massive rounds for OpenAI, Anthropic and xAI
Accel Confidential
23
Total
Funding
Valuation
2015 $40.0B $74.7B $500B
2021 $13.0B $31.7B $183B
2023 $10.0B $27.2B $75B
2024 $2.0B $2.0B $10B
2024 $2.0B $2.15B $8B
Source: Pitchbook
Note: Data as of Oct-22 2025
US - Largest AI Model Rounds 2025 Europe & Israel - Largest AI Model Rounds 2025

2022
Source: Accel analysis using State of the Art Model announcements
Pace of new model releases continues to be high
Accel Confidential
24
2023 2025
2024 2026
GPT-3.5
Text Model
Releases
Video Model
Releases
REPRESENTED COMPANIES:
REPRESENTED COMPANIES:
GPT-4 GPT-4o GPT-4o Mini GPT-4.5 GPT-5
Grok-1 Grok-1.5 Grok-2 Grok-3 Grok-4
Grok 4 Fast &
Code Fast 1
Claude 1 Claude 2 Claude 2.1 Claude 3 Family Claude 3.5 Claude 4 Claude 4.1 Opus
LLaMa 1 LLaMa 2 LLaMa 3 & 3.1 LLaMa 3.2 LLaMa 4 LLaMa 4.x/
4.5
PaLM 2
Gemini 1.0
Ultra & Pro
Gemini 1.5
Pro & Flash Gemini Nano Gemini 2.5
Pro & Flash
Gemini 3.0
Mistral 7B Mistral 8x7B Mistral 8x22B Pixtral 12B Mistral Small 3 Mistral Medium 3
Veo v1 Veo v2 Veo v3 Veo v3.1
Sora v1 Preview Sora v1 Sora 2
Kling v1 Kling 2.0 Kling 2.1 Kling 2.5 Turbo
Gen-3 Gen-4 Gen-4 Turbo
Seedance v1 PixelDance
Seaweed
Pika 2.5
PikaTwists
Pika 2.2
Pika 1.5
Pika 1.0

Source: LM Arena, ScreenSpot-Pro
LLM performance is converging, but performance gap
is still wide in video & computer-use models
25
Text LLM Performance Benchmark
Performance
Score
3% DELTA
Performance
Score
Video Model Performance Benchmark
29% DELTA
Computer Use Model Performance Benchmark
70% DELTA
0
20
40
60
80
100
H
o
l
o
1
.
5
-
7
2
B
U
I
-
T
A
R
S
-
1
.
5
7
B
Q
w
e
n
2
.
5
-
V
L
7
2
B
U
I
-
V
e
n
u
s
-
7
2
B
S
o
n
n
e
t
4
0
280
560
840
1,120
1,400
G
o
o
g
l
e
O
p
e
n
A
I
M
i
n
i
m
a
x
B
y
t
e
d
a
n
c
e
K
l
i
n
g
A
I
A
l
i
b
a
b
a
L
u
m
a
A
I
P
i
k
a
G
e
n
m
o
A
I
0
300
600
900
1,200
1,500
G
o
o
g
l
e
A
n
t
h
r
o
p
i
c
O
p
e
n
A
I
A
l
i
b
a
b
a
x
A
I
D
e
e
p
S
e
e
k
M
i
s
t
r
a
l
M
i
c
r
o
s
o
f
t
T
e
n
c
e
n
t

On the cloud and AI applications side, private EU/IL
funding performs well, representing 2/3 of the US
26
2018 2019 2020 2021 2022 2023 2024 2025E
$30B
$22B
$19B
$49B
$62B
$22B
$18B
$13B
$45B
$42B
$28B
$78B
$119B
$48B
$39B
$33B
US Cloud & AI Funding Excl. Models
EU/IL Cloud & AI Funding Excl. Models
US VS. EU/IL VENTURE CAPITAL CLOUD & AI INVESTMENTS EXCL. AI MODELS, $B
Source: Dealroom, Pitchbook, Accel analysis
39% 47% 46% 52% 63% 67% 52% 66%
EU funding as % of US

Total
Funding
Valuation
2021 $683M $1.5B $13.7B
2021 $540M $1.3B $6.0B
2023 $200M $223M $1.8B
2016 $184M $338M $2.1B
2019 $180M $253M $2.5B
Total
Funding
Valuation
2022 $900M $1.1B $9.9B
2022 $600M $1.7B $20.0B
2017 $410M $545M $6.4B
2023 $400M $738M $10.2B
2023 $350M $635M $10.0B
Source: Dealroom
Emerging winners in each geography are raising similar
sized rounds
Accel Confidential
27
US - Largest Application Rounds 2025 Europe & Israel - Largest Application Rounds 2025

Cybersecurity
2025 Total Category Funding: $2.7B
Customer Service
Productivity
Finance
Developer Tools
Media
Legal
Healthcare & Life Sciences
A new landscape of AI native applications is emerging…
28

Source: TechCrunch, LinkedIn
…with unprecedented bottoms-up adoption…
29
• AI is collapsing the builder stack. Self-serve tools are shifting
developers and creators from coding and wiring systems to
specifying intent, compressing the distance between idea and
product
• Distribution advantages are compounding. Viral adoption
through open communities and social media creates massive user
bases quickly — giving these platforms a wedge into enterprise
without expensive sales motions
• These tools are becoming new systems of record and are
critical to daily work. The outcome is unprecedented growth
$100M ARR in 8 months
$500M ARR as of Jun-25 in 30 months
10x YoY revenue growth
$200M ARR (2x in 10 months)
Adoption of AI Coding Assistants by Developers
Percentage
of
Developers
Using
AI
(%)
0
25
50
75
100
2023 2024 2025
90%
63%
36%

…and a level of efficiency never seen before in software
30
ARR ($M) / FTE
0.54
0.49
0.51
0.46
1.2
1.5
3.4
6.1

Source: Accel Analysis, CapIQ
Notes: Globalscape Average as of Oct-22 2025
Gross margin is still below traditional software…
31
GROSS MARGIN - EMERGING AI APPLICATION LEADERS (ANONYMIZED)
0
0.2
0.4
0.6
0.8
Company 1 Company 2 Company 3 Company 4 Company 5 Company 6 Company 7 Globalscape Public Cloud
Index Companies (LTM)
76%
40%
40%
33%
32%
15%
10%
7%

`
Source: Accel analysis using OpenAI announcements
Note: Data as of Oct-22, 2025
…but lowering inference cost should drive
margin expansion in the future
Accel Confidential
32
INFERENCE COSTS PER 1K TOKENS
Date of Price Change / Latest Price
03/23 11/23 05/24 10/25 10/25
$2
$10
$10
$19
$75
GPT 4 Turbo
GPT 4 GPT 4o
2x faster, half the
price, and 5x
higher rate limits
than GPT-4 Turbo
INFERENCE COSTS PER 1M TOKENS
-97% Cost Decline
in 31 months
GPT 5 GPT 5 Mini

2025
|
Globalscape
NOVEMBER 2025
GLOBALSCAPE
01
Introduction
02
03
04
05
What’s next
06
Appendix

Powering the AI industrial revolution
requires new infrastructure…
AI AS THE 21ST CENTURY UTILITY
34
+
Massive need for AI infrastructure build

…driving a race for compute
35
AI AS THE 21ST CENTURY UTILITY GLOBALSCAPE REPORT | NOVEMBER 2025

US Japan Germany France Italy UK Spain
31
36
36
53
58
116
502
2025 2026 2027 2028 2029 2030
169
135
113
92
71
52
Training
Inference
Source: Accel analysis; Mizuho - Amazon.com, Inc. (AMZN), 29th Sept 2025; Ember Energy - Yearly Electricity Data
Build out of AI data center capacity is forecasted
to reach 117 additional GW by 2030
36
Estimated Global AI Data
Center Capacity Demand
ESTIMATED GLOBAL DATA CENTER CAPACITY ENERGY DEMAND - GIGAWATTS
Electricity Consumption
by Country
ELECTRICITY CONSUMPTION OF SELECT COUNTRIES, 2024 - GIGAWATTS
103GW Combined
Electricity Consumption
Enough to power Italy, Spain and the UK combined
+117 GW / 3.3x
50%
50%
55%
60%
63%
66%
69%
45% 40%
37% 34%
31%

Source: Axios - ‘Sam Altman: OpenAI wants to get to $1 trillion a year in infrastructure spend’; Globenewswire - ‘Brookfield and Microsoft Collaborating to Deliver Over
10.5 GW of New Renewable Power Capacity Globally’; The Information - ‘Meta Plans to Spend $600B Through 2028 in the US’
55%+ of forecasted capacity by 2030 is
already committed / announced
37
(Stargate)
~65GW Combined Capacity Announced by 2030
Represents ~55% of estimated CapEx
Other
Commitments
“We have currently committed to about 30 gigawatts of compute,
with a total cost of ownership … of about $1.4 trillion”
Includes 10.5 GW renewables
agreement with Brookfield
$600B US Data Center
Spending Plan by 2028
Note: several large players (e.g. Amazon,
Google) have not made significant, specific,
public data center commitments
COMMITTED/ANNOUNCED CAPACITY (GW)
30.0GW 17.3GW 11.3GW 6.1GW

Source: Nvidia - ‘Introduction to NVIDIA DGC H100/H200 Systems’; ‘The Rising Costs of Training Frontier AI Model’ - Cottier et Al (2024); Turner & Townsend Data Centre
Cost Index 2024; Google - ‘Growing the internet while reducing energy consumption’
117GW incremental data center capacity translates to
~$4T of AI CapEx between ’26 and ‘30
38
2026 2027 2028 2029 2030
$166B
$114B
$98B
$102B
$95B
$343B
$236B
$204B
$212B
$197B
$663B
$455B
$394B
$409B
$380B
$1,172B
$804B
$696B
$723B
$671B
Compute Infrastructure (e.g., GPUs, CPUs)
Facility Build
Networking & Storage Infrastructure
Estimated AI CapEx Forecast ($B)
2026 - 2030, SEGMENTED BY COMPUTE INFRASTRUCTURE, FACILITY BUILD, AND NETWORKING & STORAGE
Incremental GW
+19 GW +21 GW
+19 GW +21 GW +20 GW +23 GW +34 GW
TOTAL
$4.1T
$0.6T (14%)
$1.2T (29%)
$2.3T (57%)
Total
CapEx

Source: RBC - ‘Telecom Services & Communications Infrastructure 10.17.2025’; CapitalIQ Estimates as of Oct-22 2025
2025 CapEx is concentrated in large hyperscalers and
their op. cash flow can finance a 5-year build out
39
Hyperscaler CapEx - 2025E ($B)
CAPEX FORECAST FOR AMZN, MSFT, GOOGL, META & AAPL
CY25E CapEx
$14B
$69B
$83B
$104B
$113B
$383B
combined
hyperscaler CapEx,
representing ~85%
of total forecast of
~$450B
Hyperscaler Operating Cash Flow vs
Estimated Total AI CapEx Forecast ($B)
2026 2027 2028 2029 2030
$1,424B
$1,257B
$1,107B
$914B
$795B
$1,172B
$804B
$696B
$723B
$671B
AI CapEx Forecast
Hyperscaler Operating Cash Flow
Note: excludes debt capacity
$14B

Source: Nebius and CoreWeave S-1s
$3.1T AI data center revenue required over 5 years to
payback ’26-’30 expected CapEx
40
Combined AI Application Revenue Est.
2026 2027 2028 2029 2030
Total AI CapEx ($B) $671B $723B $696B $804B $1,172B
Facility Depreciation ($B) - 10Y Life 20 41 61 85 119
Compute Infra Depreciation ($B) - 5Y Life 76 158 237 327 460
Networking Infra Depreciation ($B) - 5Y Life 19 39 59 82 115
Total Depreciation $115B $238B $357B $494B $694B
Annual Running Costs ($B) $40B $83B $125B $173B $243B
Annual Depr. + Running Costs ($B) $155B $321B $482B $667B $937B
Data Center Rev Required For Payback (+20% Margin) $186B $386B $578B $801B $1,125B
$3.1T Combined Data Center Revenue
Required by 2030 to payback ’26-’30 CapEx
Assumptions: Energy cost of $0.086 per kWh (YCharts - US Average Retail Price of Electricity in the Industrial Sector; PUE of
1.1; power assumed to be 40% of annual running costs for data center (University of Washington - Data Center Energy Management).
Analysis includes running costs for new capacity only over the period (i.e., excl. installed base pre-2026)

Source: IMF - GDP, current prices
$3.1T revenue equivalent to an increase
of 1-2% of global GDP CAGR
41
Incremental GDP Required to Payback CapEx
2026 2027 2028 2029 2030
Data Center Revenue
Required for Payback $186B $386B $578B $801B $1,125B
(=) Incremental Corporate
Profits Required
$186B $386B $578B $801B $1,125B
Implied Incremental GDP Required
(10x corporate profits)
+$1.9T +$3.9T +$5.8T +$8.0T +$11.2T
World GDP (IMF Forecast) $124T $130T $136T $143T $150T
Minimum World GDP Required
to justify AI Data Center Investments
$125T $134T $142T $151T $161T
’25-30
GDP CAGR
5.0%
’25-’30
GDP CAGR
6.5%
VS

Source: Morgan Stanley Research
Capital likely won’t be the constraint, but will the 36 GW
US electricity shortfall be the bottleneck?
42
Available US Grid Capacity US Data Center Capacity Under Construction Potential Shortfall US Data Center Power Needed '25-28
57 GW
36 GW
6 GW
15 GW
Potential Shortfall in Power for US Data Centers, ’25 - ‘28
GIGAWATTS
It would take 1,530 square
kilometers of solar panel coverage
to cover this shortfall (i.e., an
area bigger than Los Angeles) - or
35 nuclear reactors (+37% more
than the current US nuclear
reactor capacity)

2025
|
Globalscape
NOVEMBER 2025
GLOBALSCAPE
01
Introduction
02
03
04
05
What’s next
06
Appendix

First — Congratulations to this year’s EU/IL unicorns!
Accel Confidential
2025 GLOBALSCAPE
44
New European & Israeli Unicorns
Source: Dealroom, Pitchbook

DEVELOPERS / INFRASTRUCTURE
FOUNDATION MODELS
Accel 2025 Europe AI 100
Accel Confidential
45
Top 100 EU & Israel AI & Cloud companies over $1M revenue and valued at less than $1B
SECURITY
2025 GLOBALSCAPE
ROBOTICS HORIZONTAL AI APPLICATIONS
VERTICAL AI
FINTECH AI INFRASTRUCTURE

Source: Pitchbook, Accel Analysis
2025 Accel Europe AI 100 — At a glance
Accel Confidential
2025 GLOBALSCAPE
46
~10,400
TOTAL EMPLOYEES
$5.4B
TOTAL FUNDING
~4 YRS
AVG YEARS SINCE FOUNDING
104
AVERAGE FTEs
$54M
AVERAGE FUNDING
122%
AVG LTM FTE GROWTH

Congratulations to this year’s US unicorns!
Accel Confidential
2025 GLOBALSCAPE
47
Selected New US Unicorns
Source: Dealroom, Pitchbook

DEVELOPER TOOLING
AI INFRASTRUCTURE
Accel 2025 US AI 100
48
Top 100 US AI & Cloud companies valued at less than $1B
SECURITY
2025 GLOBALSCAPE
POKE
FOUNDATION MODELS / DATA AI-ENABLED SERVICES
VERTICAL AI HORIZONTAL AI APPLICATIONS
CONSUMER ROBOTICS

2025 Accel US AI 100 - At a glance
Accel Confidential
2025 GLOBALSCAPE
49
~5,200
TOTAL EMPLOYEES
$4.8B
TOTAL FUNDING
~2.4 YRS
AVG YEARS SINCE FOUNDING
52
AVERAGE FTEs
$48M
AVERAGE FUNDING
213%
AVG LTM FTE GROWTH

Globalscape winners have raised $10.3B in total
2025 GLOBALSCAPE
50
Horizontal AI Apps, Vertical AI and Developer Tooling are the most well-funded
categories
TOTAL CAPITAL RAISED BY CATEGORY ($M)
$2,178M (21%)
HORIZONTAL AI APPS
$2,169M (21%)
VERTICAL AI
$1,792M (17%)
DEVELOPER TOOLING
$1,360M (13%)
SECURITY
$944M (9%)
GEN AI FOUNDATION
MODEL / DATA $527M (5%)
FINTECH
$501M (5%)
AI INFRA
$458M (4%)
CONSUMER $241M (2%)
ROBOTICS
$82M (1%)
AI-ENABLED
SERVICES

Source: Pitchbook, LinkedIn, Harmonic, Accel Analysis
Note: Data as of Oct-22 2025, excludes categories with <5% of total funding
Foundation Models, and Horizontal / Vertical AI
companies exhibit strongest hiring velocity
2025 GLOBALSCAPE
51
DEVELOPER TOOLING AI INFRA FINTECH INFRASTRUCTURE SECURITY VERTICAL AI HORIZONTAL AI GEN AI FOUNDATION MODEL / DATA
202%
178%
173%
163%
159%
127%
96%
AVERAGE YOY FTE GROWTH BY CATEGORY (%)

65%+ of winners are 0-3 years old
2025 GLOBALSCAPE
52
US companies skew younger at 2.4 years average age, vs 4.1 years for EU/IL winners
<2 2-3 4-5 6-7 8-9 10+
7
11
16
33
85
48
YEARS FROM FOUNDING (#)

2025
|
Globalscape
NOVEMBER 2025
GLOBALSCAPE
01
Introduction
02
03
04
05
What’s next
06
Appendix

What’s Next?
Accel Confidential
54
AI-native
vertical
applications
Enterprise
agentic &
computer-use
deployment
AI Security Voice &
media for the
enterprise
Vibe coding
revolution in
the enterprise
WHAT’S NEXT
Legal
Finance
Construction

Source: PwC's AI Agent Survey, May 2025
Increased deployment of AI agents / computer-use
models in the enterprise
Accel Confidential
WHAT’S NEXT
55
• Addition of AI agents to existing platforms will increase
the complexity of workflows that can be automated
within enterprise functions (e.g. legal, sales, finance, and
HR) since agents are able to call tools & orchestrate
processes
• Adoption is starting but the probabilistic nature of LLMs is
still preventing mass deployments
• A new generation of tools is emerging to provide
guardrails and larger context access and unlock larger
deployments
AI-budgets increasing due to agentic AI
>50% 26–50% 10–25% <10% No increase Unsure
6%
6%
17%
45%
18%
8%
Over the next 12 months, how much is your team or business function
planning to increase its AI-related budget due to interest in agentic AI?

Source: Thompson Reuters, 2025 Generative AI in Professional Services Report
Rise of vertical-specific AI applications
Accel Confidential
WHAT’S NEXT
56
• New vertical AI platforms are automating industry-
specific workflows - for example, Abridge in medical
note-taking, PermitFlow in construction, and Harvey in
legal research
• As the capabilities of vertical-specific AI tools increase,
these systems will likely capture spend from human-
delivered services in addition to traditional software
budgets
• Industries and use cases with large documentations will
experience biggest disruption
Healthcare / Life Sciences Construction
Finance
Legal
Example in Real Estate
PermitFlow bringing automation at scale to the construction industry

Source: Team8 CISO Village Survey, July 2025
AI Security as the new cyber battleground
Accel Confidential
WHAT’S NEXT
57
• The rise of AI models and agents creates new attack
surfaces (prompt injection, data exfiltration, model
poisoning, and unauthorized tool use), posing challenges
beyond traditional cybersecurity
• Securing AI requires unified controls for data, identity,
permissioning, and observability to keep agent
behaviour trustworthy and auditable
• AI is also powering a new generation of security tools
that automate SOC workflows, boosting detection
accuracy and response speed
AI Security is now a top priority for CISOs
S
e
c
u
r
i
n
g
A
I
A
g
e
n
t
s
G
o
v
e
r
n
i
n
g
A
I
U
s
a
g
e
D
L
P
A
I
-
A
u
g
m
e
n
t
e
d
S
O
C
/
S
m
a
r
t
A
u
t
o
m
a
t
i
o
n
I
n
f
r
a
V
u
l
n
M
g
m
t
3
r
d
P
a
r
t
y
R
i
s
k
M
a
n
a
g
e
m
e
n
t
S
e
c
u
r
e
P
r
o
d
u
c
t
s
b
y
D
e
s
i
g
n
H
u
m
a
n
R
i
s
k
M
g
m
t
S
e
c
u
r
i
n
g
A
I
-
G
e
n
e
r
a
t
e
d
C
o
d
e
20%
21%
21%
23%
24%
28%
33%
36%
39%
2025 Top CISO Painpoints

Source: Stack Overflow - 2025 Developer Survey
Vibe coding is moving to the enterprise
Accel Confidential
WHAT’S NEXT
58
• AI-driven code generation is reshaping how enterprises
build, deploy, and maintain software
• Initial adoption is focused at the developer level but with
models becoming more performant, we can expect that
larger amount of code and eventually applications will
be built by agents
• Enterprise will need to rethink their development, CI/CD
and deployment cycles to embrace AI technologies
Professional developers’ usage of AI tools in
the software development process
Daily Weekly Monthly Plan to use soon No plans to use
15%
5%
13%
17%
51%

Source: Synthesia - ‘Introducing Synthesia 3.0’
AI-driven voice & media becoming
the next enterprise UX frontier
Accel Confidential
WHAT’S NEXT
59
Synthesia launching ultra realistic avatars and
video agents
• AI-driven voice and media are emerging as the next
UX frontier in the enterprise, enabling richer, more
interactive experiences at a fraction of historical cost
• Videos are now ultra realistic and interactive. First
video agents are launched
• Adoption is accelerating across customer support,
sales, onboarding, and training as enterprises
integrate synthetic media into everyday
communication and engagement.

2025
|
Globalscape
NOVEMBER 2025
GLOBALSCAPE
01
Introduction
02
03
04
05
What’s next
06
Appendix

List of companies in Globalscape Public Cloud Index
Accel Confidential
APPENDIX
61
Global Accel Globalscape Public Cloud Index — Currently Listed Companies
EU & Israel US
JFrog Israel 8x8 CrowdStrike Netskope ServiceTitan
Monday Israel Adobe CS Disco Okta Snowflake
SentinelOne Israel Amplitude Datadog Palantir Sprinklr
Wix.com Israel AppFolio Definitive Healthcare Paycom Tenable
Elastic Netherlands Asana DigitalOcean Paylocity Toast
UiPath Romania BILL DocuSign PayPal Twilio
GitLab Ukraine BlackLine Domo Procore Veeva Systems
Box Dropbox Q2 Workday
Braze Fastly Qualys Workiva
C3.ai Figma RingCentral Yext
Clearwater Analytics HubSpot Rubrik Zoom
Cloudflare MongoDB Salesforce Zscaler
Confluent nCino ServiceNow

2018 2019 2020 2021 2022 2023 2024 2025E
$8.3B
$5.0B
$3.5B
$11.2B
$12.9B
$4.5B
$3.8B
$3.2B
$1.6B
$1.7B
$1.8B
$3.1B
$4.7B
$1.8B
$1.3B
$1.3B
$4.0B
$3.1B
$2.8B
$6.5B
$7.9B
$2.0B
$2.0B
$1.3B
$4.7B
$3.2B
$3.0B
$7.0B
$6.7B
$2.5B
$2.1B
$1.8B
$7.9B
$5.3B
$4.5B
$11.7B
$12.9B
$5.4B
$4.9B
$2.4B
$6.6B
$6.0B
$4.4B
$10.0B
$16.8B
$5.7B
$4.2B
$2.8B
$33.1B
$24.3B
$19.9B
$49.5B
$61.8B
$22.1B
$18.3B
$12.9B
Israel
UK
France
Germany
Nordics
Other
European & Israeli Venture Funding
Accel Confidential
APPENDIX
62
TOTAL EU/IL VENTURE CAPITAL INVESTMENT IN CLOUD & AI BY REGION, $B

European Venture Funding
Accel Confidential
APPENDIX
63
TOTAL EUROPEAN VENTURE CAPITAL INVESTMENT IN CLOUD & AI BY REGION, $B
2018 2019 2020 2021 2022 2023 2024 2025E
$8.3B
$5.0B
$3.5B
$11.2B
$12.9B
$4.5B
$3.8B
$3.2B
$1.6B
$1.7B
$1.8B
$3.1B
$4.7B
$1.8B
$1.3B
$1.3B
$4.0B
$3.1B
$2.8B
$6.5B
$7.9B
$2.0B
$2.0B
$1.3B
$4.7B
$3.2B
$3.0B
$7.0B
$6.7B
$2.5B
$2.1B
$1.8B
$7.9B
$5.3B
$4.5B
$11.7B
$12.9B
$5.4B
$4.9B
$2.4B
$26.5B
$18.3B
$15.6B
$39.5B
$45.0B
$16.3B
$14.1B
$10.1B
UK
France
Germany
Nordics
Other

2025
|
Globalscape
Thank
you
GLOBALSCAPE
VISIT ACCEL.COM/GLOBALSCAPE FOR A COPY OF THE REPORT
NOVEMBER 2025

Accel 2025 Globalscape: Race for Compute

More Related Content

Similar to Accel 2025 Globalscape: Race for Compute

More from Razin Mustafiz

Recently uploaded

Accel 2025 Globalscape: Race for Compute