Epic uses AI to rewrite Unreal Engine ahead of UE6 | Magica
Epic says AI is rewriting Unreal Engine as it prepares UE6
Editorial Team
••📖7 min read
Epic says it is using AI to rewrite parts of Unreal Engine and plans to let developers connect their preferred models to UE6. The strategy combines claimed maintenance gains with a broader overhaul of programming, distribution and portable game assets, while rival Unity already offers outside-model connections.
Epic says AI tools are now developing Unreal Engine code, including substantial maintenance and architectural changes.
UE6 is intended to connect developers’ preferred models and combine the standalone engine with its Fortnite editor; Unity already offers model connections.
Early access is targeted for the end of 2027; reported engineering gains are company claims rather than a measured engine-wide productivity result.
Epic Games, the developer of the game Fortnite and the Unreal Engine development tools, says AI tools are writing much of the code in the engine work it describes, including substantial rendering and maintenance changes. The company is making that shift while preparing Unreal Engine 6, which promises both AI-assisted creation and a merger of its two main development environments. The test for outside studios is whether Epic’s reported coding gains produce a dependable engine and a manageable transition.
Epic Games’ roadmap depicts UE6 as the convergence of Unreal Engine and Unreal Editor for Fortnite. Source: Unreal Engine.
From coding experiments to engine changes
Unreal Engine is Epic’s game-development technology, used to build standalone games and other interactive content. Its next version will bring together UE5 and Unreal Editor for Fortnite, or UEFN, the environment for making experiences inside Fortnite. Epic’s October 2024 count put its active community at 850,000 developers building projects each month—a dated measure of the constituency affected by an engine transition, rather than a current adoption figure. The changes affect both how those studios work and how they distribute games.
Marcus Wassmer, Epic’s executive vice president of development and the development-team leader who set out its June UE6 roadmap, has since confirmed that AI is producing engine code and features. In answers supplied to questions from a games publication, he said engineers were no longer writing the bulk of the code in the work he described, but were guiding tools toward the intended architecture and iterating within memory, performance and scalability constraints. That is Epic’s account of its workflow, not a measured share of all engine code.
At Ignite, Nvidia and Microsoft are expanding local coding AI, cloud desktop streaming and Azure infrastructure. The announcements build on an existing Windows AI strategy, with AMD alternatives, staggered availability and no comparable deployment-cost results in the announcements.
Editorial Team
The chronology matters. Epic’s June roadmap said most early internal applications involved supporting work rather than generating mainline Unreal Engine code. By October, chief executive Tim Sweeney was confirming AI code development for the engine—a change from his November 2025 denial of that use, as documented in the same report. Sweeney is also Epic’s founder and wrote the original Unreal Engine, giving his endorsement a direct connection to the technology being rebuilt.
Wassmer identified AI coding tools Claude Code and Codex as prominent tools in use. He credited AI with nearly eliminating non-deterministic cooking bugs—unpredictable faults in preparing game content—that Epic had fought for a decade, and accelerating shader-compression work. He said both changes significantly reduce patch sizes.
His clearest time comparison concerned rewriting a substantial portion of the D3D12 rendering layer: a job he had estimated would take one engineer three to six weeks was completed in about one week, he said, with reduced CPU time and fewer bugs. That compares an estimate with a reported completion time; it is not a controlled measurement of how much faster Unreal development has become or a measure of net cost savings. The responses provide no engine-wide productivity figure or quantified patch-size reduction.
Epic also says automated Fortnite PC-stability monitoring is identifying problematic driver and hardware combinations, while automated diagnosis and proposed fixes are improving persistent issues. He also cited healthier and broader internal test suites. Wassmer’s broader argument is that AI makes changes across thousands of files practical and should allow more modular engine architecture and easier removal of components a particular game does not need. Those engine-wide benefits remain promises.
Epic’s UE6 roadmap envisages exposing a broad set of engine functions through the Model Context Protocol, or MCP, so developers can connect different models to engine tools and build custom integrations. Epic Developer Assistant would be an optional ready-made alternative.
The intended jobs extend beyond programming to repetitive production work: setting up levels, rigging characters, adjusting lighting and preparing other content. Epic promises to test the workflows internally and in UEFN before shipping them in UE6. This is a plan for connecting models to professional tools, not evidence that those workflows already deliver all the promised savings.
Nor is model choice through an engine connection unique to Epic. In a May 5, 2026 description of its open beta, rival engine maker Unity listed an in-editor assistant, an AI Gateway for connecting preferred tools such as Claude and GPT, and an official MCP server that gives agents project context. Unity said the tools were built into Unity 6 and powered by third-party frontier models. Its earlier Muse product used Unity’s own models and has been deprecated.
Access conditions matter, too. Unity said its Gateway lets studios use existing model subscriptions without consuming Unity AI credits. The Gateway and MCP server were included with Pro, Enterprise and Industry plans, with trial access for Personal users. The beta required a project linked to Unity Cloud and cloud connectivity. These terms describe an existing alternative, not a like-for-like cost benchmark for UE6: Epic’s roadmap does not specify comparable model costs or access terms.
Unity’s beta sets an existing baseline for Epic’s planned integrations. A meaningful comparison would need to establish how effectively each engine exposes useful operations, preserves developer control and produces reliable results.
Unity’s beta documentation shows an agent selector inside the editor. Image: Unity. Source: Unity.
A programming overhaul with ecosystem stakes
AI is one of three major strands of the UE6 plan. Another is Scene Graph, a new gameplay framework built on Epic’s Verse programming language. Verse uses transactions that can be rolled back and rerun, including when calling specially compiled C++ code.
Epic wants that mechanism eventually to distribute game logic across servers without developers writing custom networking coordination throughout their code. The distinction between current capability and ambition is substantial: the roadmap says Verse currently runs on a single thread, while distributed operation is promising prototype work. Epic also says appropriately scaling concurrency on hardware still needs work. Automatic persistence and larger shared worlds are goals of that architecture, not established UE6 production results.
The third strand is portability across games and engines. Epic proposes supporting existing formats such as glTF and USD where suitable and publishing specifications for systems that lack an adequate standard. Its first proposed demonstration is Fortnite cosmetics: an open UE6 module would let participating games use players’ entitled Fortnite Outfits and let developers make outfits usable inside Fortnite.
That connects technical interoperability to players’ purchases and developers’ distribution choices. Epic says developers will still be able to ship standalone projects, publish directly inside Fortnite or operate their own ecosystems, with compatibility with Epic’s ecosystem optional. Open specifications would make that promise broader than an integration confined to Epic’s own products; the announcement does not settle the commercial terms of its proposed shared asset economy.
There is precedent for Epic linking engine economics and distribution. In its October 2024 announcement, the company offered a reduction in engine royalties from 5% to 3.5%, starting in 2025, for eligible games released on the Epic Games Store before or alongside other stores on corresponding supported platforms. The reduced rate would apply across platforms and stores, including consoles. That historical incentive explains why studios have reason to examine both technical access and commercial terms as the UE6 ecosystem takes shape.
Epic’s company-reported 2024 terms offered eligible games a royalty reduction from 5% to 3.5%, conditional on Epic Games Store release timing on corresponding supported platforms. Chart: Magica. Source: Unreal Engine.
The migration and proof still ahead
Epic targets UE6 early access for the end of 2027, followed by a full release 12–18 months later. Its development stream is visible on GitHub, but the roadmap explicitly says it is not an alpha and that Verse is not yet intended for general adoption.
Existing projects are a practical constraint. Actors and Blueprints, parts of Unreal’s existing development framework, will remain in early UE6 versions. Epic plans to deprecate them once the replacement is sufficiently mature and provide conversion tools. UE5 changes will flow into UE6, but UE6 changes will not flow back; a further UE5 release after 5.8 remains an option rather than a commitment.
Wassmer’s statement on current development puts almost all of it in streams attached to Fortnite, UEFN and UE6. He says a possible UE5.9 would be modest and include AI-developed changes. For studios deciding when to migrate, the evidence still needed is release quality, reproducible performance and patch-size results, and working conversion tools. Those tests can turn Epic’s internal accounts of faster coding into a reason to adopt UE6.
Microsoft and Nvidia are expanding Azure’s Vera Rubin virtual-machine preview, with systems already running in its data centers. Microsoft’s separate confidential-AI showcase covers four GPUs, while the announcements leave pricing, regions and general availability unspecified.
Meta is reconsidering its 2030 climate deadline after net emissions reached 11.8 million tonnes in 2025. Hardware and infrastructure purchases dominate its reported footprint, while new gas generation adds a longer-term challenge.
Three House committee chairs want the Remote Access Security Act in the annual defense bill as scrutiny of Nvidia intensifies. The available Senate text would authorize controls on specified foreign users’ remote access, with computing limits left to Commerce.
Google Quantum AI reports tests of quantum contextuality spanning up to 105 qubits on Willow processors. The experiments expose hardware errors and beat restricted classical strategies, while the largest task relies on an unproven classical benchmark.
A bipartisan House disclosure shows requests to sell 3.4 million AI chips to China and approvals for more than 879,000. The oversight question is whether Commerce verified the buyers, supply protections and access restrictions behind those authorizations.
OpenAI says its internal Performance Factory tracked more than 1,700 investigations in just over three months. About 400 cases reached proposed fixes, but far fewer shipped repairs used Codex’s code unchanged, and engineers retain approval.
Activists protested Anthropic’s stock-market debut outside Nasdaq, with at least one arrest and conflicting accounts of the total. Their campaign targets retirement-fund exposure to AI shares, alongside concerns about jobs, data centers and public accountability.
Anthropic says cash, growing revenue and multiyear payment schedules support its vast compute commitments. Its filing shows why flexibility matters: short-term capacity costs more, while major cloud contracts require payment regardless of usage.
OpenAI is rolling out live transcription for dictated ChatGPT prompts across free and paid plans. Users can see words as they speak and review the text before sending, but the company has not supplied performance measurements or a replacement recording maximum.
Boltz has raised a Bessemer-led $110 million Series A to expand molecular-design tools sold by usage. Its open-source research brought broad adoption; the commercial test is whether paid computation delivers repeatable gains in laboratory discovery.
Donald Trump said he and Xi Jinping supported minimal AI guardrails but had not discussed definitions. Their Shenzhen meeting leaves unresolved how bilateral safety cooperation would coexist with U.S. opposition to global regulation.
Gemini can now click and type across Mac apps and websites in a background window, Google says. The paid rollout expands its desktop automation, but app permissions expose active accounts and file changes that backups cannot always undo.
The House AI Safety and Innovation Caucus promises legislative advocacy and rapid incident briefings. Its members already have bipartisan safety proposals in play, making the choice of safeguards its next test.
Binance Wallet says it will return all subscription commitments, distribute roughly 2,957 Anthropic tokenized shares free and add a $1 million token giveaway after demand far exceeded its allocation.
Early Anthropic investor Dustin Moskovitz is urging frontier AI developers to slow down. He and Cari Tuna moved their stake to a nonprofit in early 2025, directing potential returns toward philanthropy while leaving questions about the holding's value and governance rights.
A national study finds higher odds of hosting data centers in majority-Black, Hispanic and Asian neighborhoods. Facility types, zoning rules and unequal access to public hearings complicate who bears the costs and receives the benefits.
A debut-day estimate values the Anthropic shares FTX sold in 2024 at $49.7 billion. The comparison exposes the scale of subsequent paper gains, but contrasts cash raised for bankruptcy repayments with a hypothetical stock-market valuation.
Anthropic raised $81.3 billion in a record IPO and opened 34% above its offer price before retreating. Rapid revenue growth and adjusted operating profits sit alongside vast computing commitments and limited investor control.
The Kimi developer reportedly abandoned planned Nvidia B300 capacity for 2027 after new US restrictions. The deal exposes financing risks for overseas AI clouds, but neither a model delay nor a $150 million cash loss is established.