A free preview, generated by MasteryLoop HQ.
In 2021, the Texas Comptroller adopted a rule interpreting the state's R&D credit narrowly. The rule states plainly: "A design is not a business component because a design is not a product, process, computer software, technique, formula, or invention." Two of the Comptroller's own worked examples show how far this went: one scenario involving a test pile built to verify a foundation design was found not to be a "process of experimentation," and another involving computer-aided simulation and modeling for electrical system design was found not to qualify without actual experimentation. The Comptroller applied the rule retroactively to franchise tax reports due on or after January 1, 2014 — a seven-year lookback that drew industry pushback and was challenged in litigation.
SB 2206, effective January 1, 2026, rewrites the credit from the ground up. Instead of Texas writing its own definition of qualifying research, the law ties "Texas qualified research expenses" directly to line 48 of federal Form 6765 — whatever a firm reports there for federal purposes now flows through to the Texas credit for the portion attributable to research conducted in Texas. The standard rate jumps from 5% to 8.722%, with a higher 10.903% rate for research done in partnership with a Texas public or private institution of higher education; firms with no qualifying Texas research in the prior three periods instead get a flat 4.361% rate (5.451% if also partnered with a Texas higher-ed institution). The credit becomes permanent — the prior version was set to expire at the end of 2026 — and carryforward extends to 20 consecutive report periods.
Tying to the federal definition doesn't mean every design decision qualifies. Eligibility still runs through the same four-part federal test: the work must serve a permitted purpose aimed at a new or improved function, there must be genuine technical uncertainty at the outset, the firm must systematically evaluate alternatives through modeling, analysis, simulation, or testing, and the work must fundamentally rely on engineering or physical science.
In engineering and architecture work specifically, qualifying activities include structural, mechanical, electrical, and plumbing design that requires non-obvious solutions; evaluating design alternatives through modeling; energy modeling and building performance analysis; constructability and means-and-methods development for non-standard approaches; BIM clash detection that drives real design changes; sustainable and code-compliance design requiring technical analysis; site and civil engineering for unusual site conditions; and renovation or adaptive reuse work where existing conditions are unknown going in. Texas publishes no official list of qualifying activities — the Comptroller's rule points to the federal R&D credit definition instead — so these are illustrations of how that definition lands on design work, not a state-sanctioned list. What doesn't qualify: applying a standard detail the firm has used before, purely aesthetic decisions, routine code checking, and producing construction documents once the technical questions are already resolved.
A common assumption is that work paid for by a client under contract can't generate a credit, since the firm isn't funding its own research. Two factors can preserve eligibility anyway: if the firm's fee depends on meeting defined technical performance criteria rather than being paid regardless of outcome, and if the firm retains the right to reuse the details, models, or methods it develops on other projects. Contract language — performance criteria, re-performance obligations, reuse rights — directly affects whether a claim holds up.
The credit is incremental — it only applies to qualifying expenses above a base amount, calculated as 50% of the firm's average qualifying expenses over the prior three tax periods. Because the definition of what counts changed with SB 2206, firms have to go back and restate those prior-year numbers using the new federal-tied definition, not the old Texas-specific one. The risk: a firm that has no federal Form 6765 on file for those base years has nothing on line 48 to restate, which means the base may be zero — pushing the firm onto the mandatory 4.361% flat rate applied to its entire current-year Texas research spend, instead of the 8.722% standard rate on just the incremental amount.
Some firms can receive the credit as an actual cash refund rather than just an offset against franchise tax owed — specifically, qualifying new veteran-owned businesses, entities whose computed franchise tax comes to less than $1,000, and businesses whose annualized revenue falls at or below the no-tax-due threshold. Claiming it requires filing Form 05-183, due November 15 of the report year, with no extensions accepted.
The new rules apply starting with the 2026 report year — but report years due before January 1, 2026 remain governed by the old rules, including the design and service exclusions, and those open audit years don't disappear just because the law changed going forward. Firms owned through an ESOP must now file federal Form 6765 to claim the Texas credit even though the federal credit itself is unusable at the entity level, since the Texas franchise tax credit works differently for ESOP structures. Combined groups face ambiguity where membership changed during the base years, since the statute doesn't replace language that used to address that scenario. And as of this article's sourcing, the Texas Comptroller had not yet adopted a formal rule implementing the new credit.
For engineering and architecture firms, SB 2206 doesn't just raise a rate — it reverses an exclusion that kept design work out of the R&D credit entirely. A 2021 rule said as much. It stood for five years. But eligibility now runs through the same four-part federal test everyone else uses, which means the real work shifts from "can design firms claim this at all" to "can this specific firm prove technical uncertainty, document its base period, and file on time."
In your own words: why could a firm with genuinely qualifying 2026 research still end up stuck at the 4.361% rate instead of the 8.722% standard rate — and what would it need on file to avoid that?
Reading a good explanation feels like understanding it. Usually it isn't the same thing — and you don't find out which one you've got until someone asks you to explain it back.