Supply chain
From sand to model
What has to be built, in order, before anyone can train an AI model — 12 steps, and where they narrow to a chokepoint.
5 3 1 3
Raw materials for Critical materials
The materials at the start of the chain: rare earths for the magnets that move hard-drive arms and spin cooling motors, and the polished silicon wafers every chip starts as.
The record Ga: China 99% · heavy-REE magnet controls in force · He: watchlist
99 % China share, primary low-purity gallium USGS MCS 2026, gallium (no range — disclosed figure) pubs.usgs.govGlobalWafers sells to TSMC · Samsung
2 deals at this step newest first
- Jul 9, 2026 updated Nov 19
- Jul 10, 2025 updated Nov 21
Chipmaking machines for EUV lithography
The machines that print a chip's circuits. The ones that print the finest features — EUV lithography — come from a single company.
The record ASML — sole supplier
ASML sells to TSMC · Samsung · Intel · SK Hynix · Micron
No deal in the record sits at this step yet.
Chip factories
Where the chips are made: nearly every advanced AI chip, whoever designs it, is physically made in TSMC's fabs in Taiwan. Intel is one of the few companies that both designs chips and makes them in its own factories, a business it runs as Intel Foundry. Samsung, too, makes chips that other companies design, Tesla's AI chips among them.
Samsung sells to Penguin Solutions
1 deal at this step newest first
Memory for HBM memory
The high-bandwidth memory stacked beside every AI chip, which only three companies make, and the ordinary DRAM that makers such as Nanya also produce.
The record 3 makers make virtually all — SK ~56–63% rev share (IDC) · ~$11.4B/qtr · Samsung's ready capacity booked, Micron's 2026 supply agreed · HBM eats ~22% of DRAM wafers
11.4 $B/qtr (10.26–12.54) Epoch AI Chip Components hub, 2026-05 epoch.aiIn gigawatts17.3–24.7 GW/yr at the wall it could feed · Q4 2025, annualized · 3% of supply booked to other products
0–35 GW a year
How this is counted
From 9.2 GW a year of chip power that HBM could feed (8.3–10.1), Q4 2025 annualized; 3% of the quarter's supply is booked to other products.
The bar runs 0–35 GW a year at the wall, the axis the chip, HBM and packaging steps share.
GW a year at the wall of accelerators that a year of HBM output could feed. The quarter's whole HBM supply × the accelerator TDP that its four tracked buyers (Nvidia, AMD, Google, Amazon) shipped per HBM dollar that quarter, annualized, then to the wall at ×2.09–2.44. Epoch assigns nearly all of the quarter's HBM to those buyers' chips, so this is the accelerators' number seen from the memory side: it shows where the stage binds, not a ceiling with room to spare.
source: Epoch AI chip-components and chip-sales hubs × Epoch's data-center power chain epoch.ai
SK Hynix Samsung Micron Nanya Technology
SK Hynix sells to Penguin Solutions
Samsung sells to Penguin Solutions
Micron sells to Penguin Solutions
6 deals at this step newest first
- Jul 24, 2026 updated Nov 21
- Jun 2026 updated Nov 5
- Mar 25, 2026 updated Nov 21
- Mar 18, 2026 updated Nov 20
- Dec 2025 updated Nov 21
- Oct 1, 2025 updated Nov 19
Advanced packaging
The step that bonds a chip to its memory stacks — at TSMC, the CoWoS process — on substrates that need Ajinomoto's ABF film and boards from makers like Ibiden and Kinsus; outside firms such as Amkor and SPIL package and test chips for other companies, KYEC tests them, and GlobalFoundries has agreed to make silicon interposers for CoWoS at its plant in New York.
The record TSMC CoWoS ~198k wafers/qtr · Nvidia ~59% · ABF: Ajinomoto >95%
197,813 wafers/qtr (183,750–211,875) Epoch AI Chip Components hub, 2026-05 epoch.aiIn gigawatts20.5–27.5 GW/yr at the wall it could carry · Q4 2025, annualized · 16% of supply booked to other products
0–35 GW a year
How this is counted
From 10.5 GW a year of chip power that CoWoS could carry (9.8–11.3), Q4 2025 annualized; 16% of the quarter's supply is booked to other products.
The bar runs 0–35 GW a year at the wall, the axis the chip, HBM and packaging steps share.
GW a year at the wall of accelerators that a year of CoWoS output could carry. The quarter's whole CoWoS supply × the accelerator TDP that its four tracked users (Nvidia, AMD, Google, Amazon) shipped per wafer that quarter, annualized, then to the wall at ×2.09–2.44. Epoch books the rest of the wafers to 'Other', products it doesn't split by designer, so this is a ceiling only if every wafer went to accelerators.
source: Epoch AI chip-components and chip-sales hubs × Epoch's data-center power chain epoch.ai
TSMC Ajinomoto Ibiden Kinsus Amkor SPIL KYEC GlobalFoundries
Amkor sells to TSMC · Qualcomm
7 deals at this step newest first
- Nov 17, 2026
- Nov 17, 2026
- Oct 13, 2026
- Oct 8, 2026 updated Nov 20
- Jul 23, 2026 updated Nov 18
- May 2026 updated Nov 19
- Mar 16, 2026
AI chips for Accelerators
The processors AI runs on. Nvidia's GPUs dominate and AMD is the only large GPU rival; Broadcom co-designs custom chips for buyers routing around Nvidia, Marvell designs them for the cloud giants, Qualcomm has begun to for Amazon, and MediaTek has built its first AI accelerator for an American cloud company; Google, Amazon, Cerebras and Groq have designs of their own, and Taalas, which AMD agreed to buy in 2026, builds one AI model into each chip.
The record Nvidia dominant
In gigawatts19.2–34.9 GW/yr at the wall · last four quarters, to Sep 2026 at the latest
0–35 GW a year
How this is counted
From 11.5 GW a year of chip power shipped (9.2–14.3) in each maker's last four complete quarters, to Sep 2026 at the latest; 14.5 GW a year at the latest quarter's pace.
The bar runs 0–35 GW a year at the wall, the axis the chip, HBM and packaging steps share.
GW a year at the wall for the AI accelerators that the six makers Epoch tracks shipped in a year. Epoch's count of accelerator TDP shipped by Nvidia, Google, AMD, Amazon, Huawei and Cambricon over each maker's last four complete quarters: Nvidia's to September 2026, the other five's to June 2026, each maker's latest complete quarter in Epoch's data. Chips from makers Epoch doesn't track are not in it. To the wall with Epoch's own chain: ×1.53 for the rest of the server, ×1.14 for networking and other IT, then peak PUE of 1.2 (hyperscalers) to 1.4, which makes ×2.09–2.44. The band runs from Epoch's P5 at ×2.09 to its P95 at ×2.44; P5s and P95s are summed across makers, which widens it.
source: Epoch AI chip-sales hub (accelerator TDP shipped) × Epoch's data-center power chain epoch.ai
Nvidia AMD Broadcom Marvell Qualcomm MediaTek Google Trainium2/3 Cerebras Groq Taalas
21 deals at this step newest first
- Nov 17, 2026 updated Nov 21
- Sep 8, 2026 updated Nov 20
- Aug 6, 2026 updated Nov 18
- Jul 24, 2026 updated Nov 21
- Jul 22, 2026 updated Nov 20
- Apr 19, 2026 updated Nov 21
- Apr 14, 2026 updated Nov 21
- Feb 24, 2026 updated Nov 21
- Jan 14, 2026 updated Nov 19
- Dec 24, 2025 updated Oct 8
- Nov 21, 2025 updated Nov 19
- Oct 29, 2025 updated Nov 18
- Oct 23, 2025 updated Nov 21
- Oct 13, 2025 updated Nov 19
- Oct 6, 2025 updated Nov 21
- Sep 22, 2025 updated Nov 18
- Sep 4, 2025 updated Nov 21
- Mar 7, 2025 updated Nov 20
- Feb 2025 updated Nov 21
- Dec 2, 2024 updated Nov 19
- Sep 30, 2024 updated Nov 19
Networking & optics
The switches, optical modules, lasers and fiber that wire thousands of chips into one cluster.
The record Broadcom/Nvidia fabrics · 200G EML lasers the pinch
Broadcom Nvidia Marvell Qualcomm Arista Cisco Nokia Lumentum Coherent Corning InnoLight Applied Optoelectronics Astera Labs Fabrinet
Corning sells to Meta · Amazon / AWS
11 deals at this step newest first
- Oct 28, 2026 updated Nov 21
- Sep 8, 2026 updated Nov 20
- May 6, 2026 updated Nov 18
- Mar 2, 2026 updated Nov 19
- Feb 2026 updated Nov 5
- Mar 18, 2025 updated Nov 19
- Mar 13, 2025 updated Nov 21
- Mar 13, 2025 updated Nov 20
- Dec 2, 2024 updated Nov 19
- Feb 20, 2024 updated Nov 18
- Feb 23, 2023 updated Nov 21
Servers, storage & cooling
The clusters the chips are built into, the hard drives and flash storage that hold the data, and the power-distribution and cooling gear inside each hall.
Penguin Solutions Dell Supermicro Seagate Western Digital Sandisk Backblaze Vertiv Fabrinet
7 deals at this step newest first
- Oct 27, 2026 updated Nov 21
- Jun 23, 2026 updated Nov 18
- May 7, 2026 updated Nov 21
- Jan 29, 2026 updated Nov 18
- Mar 13, 2025 updated Nov 20
- Jul 2024 updated Nov 19
- Jun 21, 2024 updated Nov 18
Power for Power equipment & grid
The electricity the data centers run on: new gas turbines, some of them rented to run a data center off the grid, existing nuclear plants and small reactors still being built, fuel cells that make power on site, geothermal plants still to be built for a data-center customer — and the backup generators that keep a data center running if its grid power fails.
The record GEV turbines sold out through 2028 · <10 GW of 2029 slots · transformers 2–4 yr · queues >5 yr
122 GW turbine backlog+SRAs GEV 3Q'26 8-K ex-99.1 (no range — disclosed figure; see pw-gev-sra) sec.govIn gigawatts15.3–20.0 GW/yr of GE Vernova gas turbines, from shipped to target
How this is counted
GW a year of GE Vernova gas turbines, from what it shipped in 2025 to the annual output it targets from 3Q26. Low is what GE Vernova shipped in 2025. High is the annual output it says it is 'on track' to reach in the third quarter of 2026. That is a target neither 3Q26 filing restates: shipments were 4 GW, 3 GW and 5 GW in the first three quarters of 2026 (the releases' whole-GW figures), and 20 GW a year would be 5 GW a quarter. At its April call about a fifth of the gigawatts it had under contract were explicitly for data centers. One maker of three — Siemens Energy, which has its own record, and Mitsubishi Power are not counted here — and gas turbines are only one route to power.
From the record
- shipped in 2025 — 15.3 GW
- shipped in 1Q26 — 4 GW; the record: “shipped 4 GW of equipment”
- shipped in 2Q26 — 3 GW; the record: “shipped 3 GW of equipment”
- shipped in 3Q26 — 5 GW; the record: “shipped 5 GW of equipment”
- the annual output it is 'on track' to reach in 3Q26, a target (then 24 GW in 2028, 30 GW in 2030) — 20 GW; the record: “on track to deliver 20 GW of annual gas turbine output in the third quarter of 2026”
- the share of GW under contract in April explicitly for data centers — “the remaining 20% explicitly supporting data centers”
source: the high end, the target, from GE Vernova's 2Q'26 release (linked here); the low end from its 2025 10-K, and what shipped since from its 1Q–3Q'26 releases, each linked from its own line in pw-gev-sra sec.gov
GE Vernova Siemens Energy Generac Bloom Energy Solaris Energy Infrastructure Stateline Power Constellation Talen Energy NextEra Energy Vistra Entergy Louisiana Entergy Arkansas Georgia Power DTE Electric NiSource AES Indiana Xcel Energy NV Energy AEP Black Hills Corp. OG&E Fervo Energy Kairos Power X-energy TerraPower Oklo Ormat Technologies Quaise Energy
28 deals at this step newest first
- Nov 17, 2026 updated Nov 21
- Nov 11, 2026 updated Nov 21
- Oct 16, 2026 updated Nov 21
- Oct 14, 2026 updated Nov 19
- Oct 5, 2026 updated Nov 21
- Sep 30, 2026 updated Nov 21
- Sep 16, 2026 updated Nov 18
- Sep 1, 2026 updated Nov 21
- Jul 22, 2026 updated Nov 21
- May 7, 2026 updated Nov 18
- Apr 22, 2026 updated Nov 21
- Mar 17, 2026 updated Nov 21
- Mar 2, 2026 updated Nov 21
- Feb 24, 2026 updated Nov 21
- Feb 17, 2026 updated Nov 18
- Jan 9, 2026 updated Nov 21
- Nov 20, 2025
- Oct 30, 2025 updated Nov 21
- Oct 27, 2025 updated Nov 21
- Oct 13, 2025 updated Nov 21
- Sep 5, 2025 updated Nov 21
- Jul 24, 2025 updated Nov 18
- Feb 20, 2025 updated Nov 21
- Oct 30, 2024 updated Nov 20
- Oct 16, 2024 updated Nov 19
- Oct 14, 2024 updated Nov 19
- Sep 20, 2024 updated Nov 21
- Mar 2024 updated Nov 19
Data centers
The buildings that hold the chips — built by developers and landlords, several of them former bitcoin miners, and leased to the clouds and labs.
In gigawatts19.5 GW at the frontier sites Epoch tracks (13.9–27.3, Nov 2026; 96 sites)
then 22.1 GW end-2026 (15.8–30.9) → 32.6 GW end-2027 (23.3–45.6) → 50 GW end-2028 (35.7–70.0)
How this is counted
GW of facility power at the frontier AI sites Epoch tracks worldwide. Facility power — what the whole site draws at its peak, cooling included, not its IT load — summed over every site in Epoch's hub, each at its latest timeline point on or before the date. Points after today are Epoch's projections from permits, construction and announcements. Its timelines end by 2028 for all but two sites, most of them in 2026 or 2027, and a site counts at its last point after that. The band, ×1.4 either way, is the accuracy Epoch states for one site's IT power, 80% of the time: an 80% interval, where the chip stages use P5–P95. It is applied here to facility power, and kept for the sum because many sites rest on the same models and defaults (Epoch's cooling model, its peak-PUE defaults), so their errors need not cancel.
source: Epoch AI Frontier Data Centers hub (CC-BY), fetched 2026-11-17; the band from its methodology page epoch.ai
Crusoe Core Scientific Aligned DC Digital Realty Khazna SB Energy TeraWulf Riot Platforms Hut 8 Cipher Digital CleanSpark Digi Power X Zone Frontier AIB Data Centers EverWind Fermi America Kazakhtelecom Blackstone Digital Infrastructure Trust Applied Digital TCS HyperVault WhiteFiber Blockfusion
Sites Stargate Abilene · Colossus 1 · Colossus 2 · Project Rainier · Meta Prometheus · Meta Hyperion
31 deals at this step newest first
- Nov 16, 2026 updated Nov 21
- Oct 16, 2026 updated Nov 21
- Sep 30, 2026 updated Nov 21
- Aug 10, 2026 updated Nov 21
- Aug 10, 2026
- Jul 29, 2026
- Jul 27, 2026 updated Nov 21
- Jul 6, 2026 updated Nov 20
- Jul 2026 updated Nov 21
- Jun 30, 2026
- by Jun 6, 2026 updated Nov 21
- May 6, 2026 updated Nov 19
- May 5, 2026 updated Nov 21
- May 4, 2026 updated Nov 21
- Dec 18, 2025
- Dec 17, 2025 updated Nov 20
- Dec 4, 2025
- Nov 12, 2025 updated Nov 21
- Nov 3, 2025 updated Nov 21
- Oct 28, 2025 updated Nov 21
- Oct 21, 2025 updated Nov 21
- Sep 25, 2025 updated Nov 20
- Sep 23, 2025 updated Nov 20
- Aug 14, 2025 updated Nov 20
- Aug 2025 updated Nov 21
- Jul 31, 2025 updated Nov 19
- Jun 2, 2025 updated Nov 20
- May 22, 2025 updated Nov 21
- Mar 28, 2025 updated Nov 20
- Oct 15, 2024 updated Nov 18
- Jun 21, 2024 updated Nov 18
Clouds & compute providers for Cloud & capacity
The companies that run the data centers and sell the computing — the big clouds and the newer GPU-rental 'neoclouds'.
The record hyperscalers + neoclouds
Microsoft Amazon / AWS Google Oracle CoreWeave Nebius IREN Nscale Akamai Lambda Fluidstack TensorWave RUM Group ChronoScale Boost Run Crusoe HIVE Digital SpaceX SK Telecom Bell Canada NAVER KT G42 Firebird HUMAIN
37 deals at this step newest first
- Nov 16, 2026
- Oct 30, 2026 updated Nov 21
- Oct 6, 2026 updated Nov 19
- Sep 24, 2026 updated Nov 21
- Sep 23, 2026 updated Nov 18
- Aug 31, 2026 updated Nov 19
- Aug 26, 2026 updated Nov 19
- Aug 24, 2026
- Aug 19, 2026 updated Nov 21
- Jun 22, 2026 updated Nov 18
- Jun 18, 2026 updated Nov 18
- Jun 15, 2026 updated Nov 18
- Jun 5, 2026 updated Nov 21
- May 28, 2026 updated Nov 19
- May 7, 2026 updated Nov 21
- May 6, 2026 updated Nov 21
- Apr 2026 updated Nov 20
- Apr 2026 updated Nov 19
- Nov 18, 2025 updated Nov 20
- Nov 3, 2025 updated Nov 21
- Nov 3, 2025 updated Nov 19
- Nov 3, 2025 updated Nov 19
- Nov 2025 updated Nov 21
- Oct 28, 2025 updated Nov 21
- Oct 28, 2025 updated Nov 18
- Oct 2025
- Sep 25, 2025 updated Nov 19
- Sep 16, 2025 updated Nov 20
- Sep 2025 updated Oct 7
- Sep 2025 updated Nov 21
- Sep 2025 updated Nov 19
- Jul 2, 2025 updated Nov 21
- Jun 2025 updated Nov 21
- Mar 2025 updated Nov 19
- Jun 4, 2024 updated Nov 18
- Apr 15, 2024 updated Nov 18
- Jun 1, 2023 updated Nov 20
AI labs for Frontier labs
Where the chain ends: the labs that train and run the models, and buy the computing everything above produces.
The record 10+ players
In gigawatts17.2–39.9 GW signed or in prospect for three labs
OpenAI 14.9–22.2 GW · Anthropic 1.1–10.6 GW · Meta 1.2–7.1 GW · chip deals not added: 38.5 GW announced, 3.3 GW binding
How this is counted
Chip deals, beside the sites and never added: 38.5 GW announced, 3.3 GW binding.
GW of sites and compute that OpenAI, Anthropic and Meta have signed for, or have in prospect, as our records state them. Low counts signed contracts only. High adds options, announced scope, deals not yet signed or only under a letter of intent, and contested or press figures. A site counts when a lab leases it, through a subsidiary too (Meta's at Sandersville), or even from a venture it part-owns and builds for: Meta's deals at Hyperion and El Paso, though Hyperion counts in the high end only, because the Hyperion record's one gigawatt figure is the contested expanded size. A site a lab owns outright has no contract to count, which is why xAI is not here (the next stage counts Colossus). A megawatt counts once: the compute Anthropic reportedly buys at Hut 8's Beacon Point through Lambda is not added to the optional capacity Hut 8 offered Anthropic across its pipeline, which may already include it. Nor is the 'more than 1 gigawatt' of Anthropic compute 'in Fluidstack-based sites' added, the capacity that the first $35B of the financing platform Broadcom set up with Apollo and Blackstone serves (the Anthropic–Fluidstack record): Broadcom's release calls it 'previously-announced', and it reads as the Google TPU gigawatt for 2026 counted here, though no party says they are one. No party names those sites either, but River Bend, counted firm, is dated outside that 2026 gigawatt: Hut 8 targets its first data hall for the second quarter of 2027, and its November 2026 results say construction is on track. Oracle's newer sites for OpenAI count in the high end only: their figure is an 'up to', and the parties' own texts disagree on whether a later total includes Abilene. So do Anthropic's 'up to' 5 GW of Amazon Trainium capacity and 'up to' 1 GW of Nvidia systems from its deal with Microsoft and Nvidia. That announcement also has Anthropic contract 'additional compute capacity up to one gigawatt' beside its Azure purchase without saying whether it is the same gigawatt, so one is counted. Nor does any document say whether Anthropic's Nvidia-powered rentals fill part of it: its prospectus names the Nvidia partnership together with its arrangements with SpaceXAI and other clouds. So Colossus 1, rented from SpaceX, is listed but not added, as Beacon Point is not. Critical-IT load goes to the wall at ×1.2 (low) and ×1.4 (high), Epoch's peak-PUE defaults; a grid or utility draw, or a figure that doesn't say which, counts as filed. Chip deals are listed beside the sites they would fill, never added to them. Nvidia's letter of intent with OpenAI is listed, its outcome disputed: no filing says whether it still stands, and the press reports it superseded by the $30B OpenAI counts from Nvidia in its 2026 round. Broadcom's TPUs for Anthropic from 2027 are disputed at 3.5–5 GW, and the chip total takes the 3.5 GW Broadcom filed. Not counted: contracts our records state only in dollars (OpenAI's with Microsoft, the labs' with CoreWeave, Meta's cloud rentals).
Firm — counted in the low and the high
- OpenAI · PORTS-Pike, Ohio: the grid draw behind 17 leases for 8 GW-IT — 10 GW of grid or utility draw; the record: “~10 GW of gross power”
- OpenAI · AWS Trainium capacity, ramping from 2027 — 2 GW; the record: “~2 GW of Trainium capacity”
- OpenAI · Cerebras, binding both ways, by end-2028 — 750 MW
- OpenAI · Abilene, Texas: the campus Oracle leases whole for OpenAI, 618 MW of it delivered by Sep 2026 — 1.2 GW
- OpenAI · Milam County, Texas: critical IT in two 15-year leases from SB Energy — 753 MW of IT load, ×1.2–1.4 to the wall
- Anthropic · Hawesville, Kentucky: the site's gross power, for a 20-year lease — 480 MW of grid or utility draw; the record: “up to 480 MW of gross power availability”
- Anthropic · Rockdale, Texas: critical IT, 20-year lease — 191 MW of IT load, ×1.2–1.4 to the wall
- Anthropic · Point Tupper, Nova Scotia: the first phase's grid load, on a lease from EverWind — 95 MW of grid or utility draw
- Anthropic · River Bend, Louisiana: utility capacity under Fluidstack's lease 'for Anthropic' — 330 MW of grid or utility draw
- Meta · Sandersville, Georgia: critical IT, a 20-year lease from CleanSpark to a Meta subsidiary — 175 MW of IT load, ×1.2–1.4 to the wall
- Meta · El Paso, Texas: leased whole from Meta's venture with BlackRock's funds, closed in July 2026; online from 2028 — 1 GW
Upside — counted in the high only
- OpenAI · Cerebras, at OpenAI's option by end-2030 — 1.25 GW
- OpenAI · Stargate UAE cluster, announced — 1 GW of IT load, ×1.2–1.4 to the wall
- OpenAI · Oracle's four newer sites for OpenAI, 'up to' this much on top of Abilene; customer delivery from 1H27 — 4.5 GW
- Anthropic · River Bend expansion, a right of first offer — 1,000 MW of IT load, ×1.2–1.4 to the wall; the record: “up to 1,000 MW”
- Anthropic · Hut 8 optional capacity beyond River Bend — 1,050 MW; the record: “up to 1,050 MW of 'additional optional capacity'”
- Anthropic · Google Cloud TPUs, which Anthropic 'plans to access' — 1 GW; the record: “'well over a gigawatt' online in 2026”
- Anthropic · AWS Trainium capacity, 'up to' this much, nearly 1 GW of it by the end of 2026 — 5 GW
- Anthropic · Nvidia systems under the deal with Microsoft and Nvidia, 'initially' up to this much; who hosts them is not said — 1 GW
- Meta · CleanSpark's Texas sites: a letter of intent and exclusivity, not a lease — 885 MW
- Meta · Hyperion, Louisiana: the expanded size, contested — 5 GW; the record: “Hyperion expanded to 5 GW”
Chip deals — beside the sites, never added
- OpenAI · AMD Instinct framework — 6 GW; the record: “up to 6 GW”
- OpenAI · AMD: the binding first gigawatt — 1 GW
- OpenAI · Broadcom custom accelerators — 10 GW
- OpenAI · Broadcom: the slice its CEO calls contractual, for 2027 — 1.3 GW
- OpenAI · Nvidia systems, 'at least' this much under a letter of intent whose outcome is disputed: no filing states one, and the press reports it superseded by Nvidia's $30B in OpenAI's 2026 round — 10 GW
- Anthropic · AMD Instinct — 2 GW; the record: “up to 2 GW”
- Anthropic · TPUs through Broadcom from 2027, disputed at 3.5–5 GW; the filed 3.5 is the one counted — 3.5 GW; the record: “~3.5 GW 'through Broadcom' from 2027”
- Meta · AMD Instinct framework — 6 GW; the record: “up to 6 GW”
- Meta · AMD: the binding first gigawatt — 1 GW
- Meta · MTIA with Broadcom, the initial commitment — 1 GW; the record: “>1 GW initial MTIA commitment”
From the record
- Anthropic · Colossus 1 from SpaceX, not added: it may fill part of the Nvidia gigawatt counted above, and either side may exit on 90 days' notice — 300 MW; the record: “more than 300 MW of new capacity”
source: our deal records — every part names its own
OpenAI Anthropic Google Meta SpaceX Reflection AI Poolside World Labs Thinking Machines Lab Cohere
3 deals at this step newest first
- Sep 28, 2026 updated Nov 16
- Aug 20, 2026 updated Nov 19
- Jan 24, 2025 updated Nov 18