That's not a generally accepted idea. Either you've misremembered, or the source you were reading was indeed marginal (the field of etymology, sad to say, sees a lot of cranks.)
The definition of a "regular" verb is not very exact. As time changes, some verbs that were once regular become irregular due to sound changes, and some verbs that were once irregular become regular due to analogy. The ancestor of the verb "hide" in Old English was regular, but it is irregular in modern English due to a process of vowel shortening. The ancestor of the verb "seethe" had a relatively "irregular" consonant alternation ("sodden" comes from the old past participle) but it is regular in modern English because new past forms were made using the present form as the basis. Occasionally, we even see regular verbs become "irregular" due to analogy, as with dive developing the past form dove based on drive and the like, or sneak developing the past tense snuck. Generally, irregular verbs are disproportionately common in a language, because the uncommon verbs become regularized more quickly.
So, which verbs are regular and which are irregular is not actually consistent even across related languages, although there are some shared irregular words that are extremely common, and shared regular words that are less common (which might give this impression at first glance).
Also, it's unclear what you mean by "the same words" in different languages. Due to changes in meaning, semantically equivalent verbs in different Indo-European languages are often not etymologically related to each other.
Examples comparing English to French (since that's the language I know best after English):
Irregular in both languages:
- "be" = French être. In both languages, this verb is "suppletive," which means not all of its forms come from the same source. Half of the irregularity is due to this, and actually, the pattern of suppletion is different in each language. Some parts of them are historically related, though (like English is and French est for the third person present singular form).
- "have" = French avoir. BUT note that these words are not descended from the same source.
- "go" = French aller. BUT note that these words are not descended from the same source.
- "know" = French savoir and connaître. English know is related to the middle part of connaître. The "irregularities" are mainly of unrelated origin.
Irregular in English, regular in French:
- "speak" = French parler
- "fall" = French tomber
- "leave (let alone)" = French laisser
- "sing" = French chanter
- "give" = French donner
- "find" = French trouver
Regular in English, irregular in French:
- "live" = vivre
- "die" = mourir
- "want" = vouloir
- "please" = plaire
- "follow" = suivre
To put the claims you mention in perspective, it might also help to compare irregular verbs in English and other Indo-European languages to those in Basque and Hungarian (which are not related to Indo-European languages (or to each other) within a timeframe that we know of). Wikipedia says that Basque has irregular verbs meaning things like "be," "have," and "say," and Hungarian has irregular verbs meaning "to be", "to come" and "to go." Even though they're not related, it shouldn't be surprising that the irregular verbs have similar meanings in Indo-European languages, Basque, and Hungarian, for two reasons:
- verbs meaning things like "be" or "have" are usually among the most commonly used verbs in a language, and therefore more likely to be irregular for the reason I mentioned earlier
- languages near each other can influence each other, even if they don't share a common source. For example, even though the "have" verb in Germanic languages and the "avoir" verb in Romance languages are not originally related, they are thought to have influenced each other in usage.
The closest thing I can think of: In English, many "irregular" verbs are the remnant of Germanic "strong verbs," (which generally descend from Proto-Indo-European verbal roots) while English regular verb morphology is based on Germanic "weak verbs," which was the class for verbs derived from non-verbal Proto-Indo-European roots.