Key takeaways

  • Weebit Nano lifted FY26 revenue guidance to at least A$13.5 million on 20 July 2026, its third upward revision in roughly six months, yet the shares fell on the news.
  • The company has signed six commercial agreements and counts SkyWater, DB HiTek, onsemi and Texas Instruments among its foundry and licensee relationships, but first royalty contributions are not expected until FY27.
  • About A$159 million of cash sat on the balance sheet after the March 2026 quarter following roughly A$102 million of raisings, though the proceeds are earmarked for accelerated spending.

Weebit Nano has been among the most closely followed technology names on the Australian market, with the conversation centred on how far its resistive RAM licensing story could run.

That mood shifted in July. A share price that climbed steadily through the first half of calendar 2026 surrendered much of the advance in weeks, even as the company kept nudging revenue expectations higher. Better disclosed numbers alongside a weaker share price is an unusual pairing.

Latest announcement or development

On 20 July the company told the ASX it expects FY26 revenue of at least A$13.5 million, up from at least A$12 million indicated in April and the more-than A$10 million first guided at the half-year. Management attributed the increase primarily to expansion of existing customer projects rather than fresh signings.

It was the third upward revision in six months and implies revenue more than triple the A$4.41 million recorded in FY25. Audited accounts are due 28 August 2026. The reaction is the telling part: shares fell on an upgrade, suggesting attention has moved from the headline growth rate to what sits underneath it.

Recent financial and operational performance

Half-year accounts to 31 December 2025 showed record revenue of A$5.6 million, roughly eight times the prior corresponding period, with A$82.7 million of cash at bank. The loss widened year-on-year. FY25 brought a net loss of A$38.4 million, 7 per cent better than FY24, and a loss per share of A$0.19.

Operationally the period was dense: ReRAM qualification to JEDEC standards at DB HiTek, first chips received and tested from onsemi, a Texas Instruments licence, eight newly granted patents and a United States subsidiary. On 4 May two product customers taped out designs incorporating its memory, one already demonstrating a functional prototype.

Cash receipts are lumpy. The September 2025 quarter produced a record A$7.3 million and closed with A$91.6 million of cash; the March 2026 quarter delivered A$0.8 million, taking nine-month receipts to A$16.7 million.

Business model and principal assets

Weebit does not manufacture chips. It licenses ReRAM, a non-volatile memory that stores data by altering the resistance of an oxide layer. Income arrives three ways: upfront licence fees, non-recurring engineering and technology-transfer payments recognised as projects progress, and per-wafer or per-chip royalties once a licensee reaches volume production.

The first two are real cash but finite and project-bound. The third is the annuity the equity story rests on. Principal assets are the patent estate, the productised oxygen-vacancy ReRAM process modules, and the foundry relationships: SkyWater's S130 platform, DB HiTek's 130nm BCD process, onsemi's Treo platform on a 300mm line, Texas Instruments and France's CEA-Leti.

Growth opportunities

The commercial argument rests on embedded flash becoming awkward and costly to manufacture at 28nm and below, where mask counts, thermal budgets and programming voltages work against it. Industry analysis credits ReRAM with lower wafer cost, better endurance and reduced operating voltages at those nodes, with automotive updates, secure elements and edge inference the early adopters.

Weebit has said it is working towards eight to ten foundry relationships within five years, each capable of supporting fifteen to twenty product companies. Part of the recent raising funds in-memory compute and neuromorphic research on a stated horizon of four years or more, which is optionality, not revenue.

Balance sheet, funding and cash flow

March 2026 brought an A$80 million institutional placement at A$4.05 a share, a 10.8 per cent discount to the A$4.54 close on 25 March. A further A$7 million went to Israeli investors and a share purchase plan raised about A$15 million at the same price, lifting proceeds to roughly A$102 million.

That left about A$159 million on the balance sheet after the March quarter. Against the FY25 loss of A$38.4 million the runway looks unhurried on a static view, but the money was raised to spend faster: roughly A$25 million each for commercial scale-up, AI architectures and core ReRAM research. Burn should rise, not fall. No dividend has been paid.

Principal risks

Revenue composition is the live question. Engineering and technology-transfer fees are cash today, but they do not compound; the share price has long implied a royalty stream that has not started. The company points to FY27 for first royalties, which depends on customers moving from prototype silicon to volume production, a step no licensor controls.

Beyond that sits concentration across only six agreements, the risk that the onsemi qualification targeted for end-2026 slips, and competition from MRAM, faster and entrenched in cache applications, and from incumbent flash suppliers running fully depreciated fabs. Further dilution is possible if timelines extend.