Did OpenAI’s Models Improve AMD Chip Designs By Up To 14%? What Was Tested
AIThis post was created with the assistance of artificial intelligence (AI).

🔍 Read the full analysis: Did OpenAI’s Models Improve AMD Chip Designs By Up To 14%? What Was Tested on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get hardware and tech essentials delivered free with Prime

  • Fast, free delivery on millions of items
  • Prime Video, Amazon Music and more included
  • Member-only deals all year
Start your free Prime trial Free trial for eligible customers · Cancel anytime
As an affiliate, we earn on qualifying purchases.

TL;DR

OpenAI said on November 18 that its models rewrote parts of AMD chip designs, with the best reported power, performance and area result reaching 14% against heavily optimized baselines. The reported results also include single-digit gains in selected CPU blocks, but AMD has not published its own figures and no independent audit is reported.

OpenAI said November 18 that its models rewrote portions of AMD processor designs and produced up to 14% improvements in power, performance and area (PPA) on AMD’s flagship graphics processor design. The figure is OpenAI’s reported best result, not a disclosed average: the company has not named the specific GPU blocks measured, and AMD has not published its own figures.

OpenAI’s announcement covered several separate tests, rather than one result across an entire chip. For AMD’s flagship GPU, the company reported a best-case PPA improvement of up to 14% compared with what it described as heavily optimized baselines. OpenAI said some model-generated changes were included in a recent design AMD sent for manufacturing, while others were added to design files for future products. The available reports do not say which blocks received the changes or how much of the GPU they represent.

OpenAI also said AMD verified single-digit percentage PPA gains in key blocks of its next flagship CPU. That is a distinct claim from the GPU’s up-to-14% result. AMD CEO Lisa Su praised the results on X, according to the source material, but AMD has not issued a public technical breakdown or its own quantified account.

The models worked at the register-transfer level (RTL), where engineers write code describing how circuit blocks operate. OpenAI said Synopsys design tools checked that proposed rewrites preserved the original design’s behavior and evaluated their power, performance and area trade-offs. These checks were performed by conventional tools, not by the models themselves. The reported method therefore involved model-generated alternatives followed by established verification and design software.

At a glance
reportWhen: OpenAI announced the results November 1…
The developmentOpenAI reported that its models generated changes to AMD chip designs with up to 14% improvements in power, performance and area.
Crypto market snapshot
Fear & Greed Index
10/100 — Extreme Fear
Bitcoin BTC$62,868▼ 0.4%
Ethereum ETH$1,778▲ 0.5%
Tether USDT$0.9992▼ 0.0%
BNB BNB$586.59▲ 0.5%
USDC USDC$0.9996▼ 0.0%
XRP XRP$0.9735▼ 0.7%
Solana SOL$74.14▲ 0.4%
TRON TRX$0.3106▲ 0.6%
Live data · CoinGecko · alternative.me (24h change)

AI Changes in Chip Design

If the results hold up in production, the work could show how AI tools might help chip teams explore design options faster, including in hardware used to run AI systems. OpenAI characterized the AMD GPU result as “roughly a year of progress achieved in less than 5 months.” That is the company’s characterization of the work, not an independently verified estimate of development time saved across a full chip program.

The distinction between a promising block-level result and a verified whole-chip improvement matters. A gain in one circuit block may not translate directly into the same percentage change in a finished processor, where power, timing, area, manufacturing constraints and interactions among blocks all affect the final design. OpenAI says some changes have entered a design sent for manufacturing, but the source material provides no production results or independent measurements.

The work also takes place amid a commercial relationship between the companies. OpenAI holds a warrant for up to 160 million AMD shares, granted in October 2025, that vests as AMD chips are deployed, according to the supplied source. That relationship does not establish that the reported technical results are inaccurate, but it is relevant when weighing claims that have not been independently audited.

Products Worth Considering

LG 27GS60QC-B Ultragear 27-inch Curved Gaming Monitor (QHD, 180Hz, 1ms, 1000R)

LG 27GS60QC-B Ultragear 27-inch Curved Gaming Monitor (QHD, 180Hz, 1ms, 1000R)

Logitech MX Keys S Combo Wireless Keyboard and Mouse, Graphite

Logitech MX Keys S Combo Wireless Keyboard and Mouse, Graphite

As an affiliate, we earn on qualifying purchases.

From RTL Rewrites to Silicon

The reported AMD result concerns one stage of chip development. At RTL, engineers describe a circuit’s behavior in code. Synthesis software translates that description into logic gates; later design steps arrange and connect those gates on a physical chip. OpenAI’s models proposed different implementations of existing functions, while Synopsys tools checked behavior and assessed design trade-offs, according to OpenAI as reported by ANI and other outlets.

OpenAI also described tests on its own custom inference chip, called Jalapeño. It said model-proposed changes reduced standard-cell area by about 13% in one unit and 10% in another, with the changes validated after the first chips came back. The supplied accounts differ on the name of the first unit: ANI calls it a vector unit, while Inside AI calls it a warp engine. Inside AI also reported examples involving a multiplier redesign and an adder redesign, each with about 15% less cell area in synthesis; the multiplier’s reported latency fell from two clock cycles to one, while the adder remained at two cycles.

A separate benchmark tested five OpenAI models on two open-source RISC-V processor cores, PicoRV32 and Ibex, across 9,881 scored runs, according to Inside AI. GPT-6.2 Astra scored highest in that benchmark. Those open-source-core tests are separate from the AMD claims and do not independently verify the reported GPU or CPU improvements.

“Su praised the results on X.”

— AMD CEO Lisa Su, as quoted in the supplied source

Missing Blocks and Independent Checks

OpenAI has not identified the GPU blocks behind the 14% best-case figure, their share of the processor, the typical gain across tested blocks, or the design stage at which the result was measured. The available reports also do not provide enough detail to independently reproduce the test or compare the result with a complete, independently measured chip.

It is not clear how much of the reported improvement will remain after physical design and manufacturing, or whether the changes sent for manufacturing will deliver the claimed trade-offs in finished hardware. The source material reports no independent audit. AMD’s public technical evidence is also limited: OpenAI says AMD verified single-digit gains in selected CPU blocks, but AMD has not published measurements of its own. OpenAI’s financial tie to AMD is a further reason to distinguish the company’s claims from independently confirmed results.

Evidence After Manufacturing

The next useful evidence would be technical detail from OpenAI or AMD identifying the tested blocks, baselines, measurement stage and distribution of results—not just the best case. For the design AMD sent for manufacturing, later results could clarify whether the model-generated changes survive implementation and deliver the reported gains in silicon. No such production measurements are included in the material available for this report.

AMD could also publish its own account of the CPU results and the role of the changes in its product designs. Until more detail or independent validation appears, the up-to-14% figure should be read as an OpenAI-reported result for unspecified portions of a design, rather than a confirmed improvement across an entire AMD GPU.

Key Questions

What did OpenAI say improved by up to 14%?

OpenAI reported a best-case improvement in power, performance and area for parts of AMD’s flagship GPU design. It did not identify the measured blocks or give an average across the chip.

Did AMD confirm the 14% figure?

The supplied reporting attributes the up-to-14% GPU result to OpenAI. OpenAI separately said AMD verified single-digit gains in key blocks of a next flagship CPU. AMD has not published its own figures in the material provided.

What did the models actually do?

They proposed rewrites to RTL code, which describes circuit behavior. Synopsys tools then checked whether the changes preserved that behavior and assessed design trade-offs, according to OpenAI.

Were the reported improvements independently audited?

The source material reports no independent audit of the AMD chip-design results. The specific GPU blocks, average gains and measurement stage also remain undisclosed.

Are the results already in manufactured AMD chips?

OpenAI said some model-generated changes were included in a design AMD sent for manufacturing. The available material does not report results from manufactured chips that would show whether the claimed improvements held up in production.

Source: ThorstenMeyerAI.com

Nothing in this article is financial or investment advice. Cryptocurrency and precious-metal investments carry significant risk — do your own research and consider a licensed advisor.
You May Also Like

Watch Out: Bridgeline Digital (BLIN) Is Emerging as a Hot AI Stock Poised to Skyrocket

Could Bridgeline Digital (BLIN) be the next big player in AI, with rising partnerships and innovative features that could lead to explosive growth?

Layer‑2 TVL Hits All‑Time High: What This Says About Scaling Wars

Layer‑2 TVL reaching new heights signals a pivotal shift in Ethereum scaling, raising questions about the future of blockchain scalability solutions.

Bitcoindailyupdate Reports: Sweden’s Hexagon Sees Unexpected Q4 Operating Profit Increase!

The unexpected Q4 profit surge at Sweden’s Hexagon raises questions about their strategy—what could this mean for the future of the industry?